Informal caregivers are a critical part of the long-term care system in the US. Public health surveillance of informal caregivers in the United States is necessary to promote healthy aging in the community, inform public support systems and research the economic cost of long-term care. However, the surveillance methods for studying caregivers vary widely. Published estimates of the number of informal caregivers in this country range from 3.5 million to 44.4 million. This wide variation in estimates stems largely from a lack of consensus in the field regarding the definition of caregiving. The nature, scope and impact of caregiving may be particularly underestimated in racially and ethnically diverse communities. This proposed dissertation research will review current methods used to study caregivers in national data sets and explore how differences in definition and survey design influence the estimated prevalence and profile of caregivers in the United States. To date no researcher has compared caregiving across data sets. Findings from this study will inform future research on how concept definition and survey design can affect caregiving research, and serve as a guide for the interpretation of research from current caregiver data sets. The first aim (1) is to describe methodological differences in national surveys of caregivers in the United States, including key issues in the definition of caregiving and survey design methodology. The second aim (2) is to examine the implication of survey design methodology on prevalence estimates in three major national surveys used for caregiving estimates, the National Long Term Care Survey, Health and Retirement Survey and Survey of Income and Program Participation, when definition is held constant. Particularly of interest is the impact of survey methodology on the prevalence estimates of particular subgroups within the national caregiving population (e.g. racial and ethnic minorities, young adults, and male caregivers). The third aim (3) is to investigate how different definitions of caregiver affect estimates of caregiver numbers and characteristics. [unreadable] [unreadable] [unreadable] [unreadable]